godz.online
Back to tools

TLDR reference

Well-known ports

A searchable reference for the network ports you actually meet, each with its service and transport. Type to search by number or service, or filter by purpose. Everything runs in your browser.

45 shown

How it works

A searchable reference for the network ports you actually meet, each mapped to its service, transport, and a one-line note. A port is a number that lets one machine run many services at once, so the operating system knows which program a packet belongs to: web traffic to 443, mail to 587, a Postgres database to 5432. Type to search by number or by service - "5432" or "postgres" both land on the same card - or use the chips to browse just the databases, just the email ports, and so on.

It covers the classic well-known ports below 1024 alongside the commonly used higher ports that developers care about day to day, such as 3306 for MySQL, 6379 for Redis, and 27017 for MongoDB. Each entry notes whether the service runs over TCP, UDP, or both. Everything is static and runs in your browser, so the lookup is instant and works offline once the page has loaded.

Example. Searching "443" shows HTTPS over TCP, the default for the encrypted web, while filtering by the Database chip lines up 3306 (MySQL), 5432 (PostgreSQL), 6379 (Redis), and 27017 (MongoDB) so you can see the defaults side by side when wiring up a connection string.

FAQ

What are well-known ports?

Ports are split into three ranges. The well-known ports, 0 to 1023, are reserved for core services such as HTTP (80), HTTPS (443), and SSH (22), and on most systems only an administrator may bind to them. The registered ports, 1024 to 49151, are assigned to specific applications such as MySQL (3306) and PostgreSQL (5432). The dynamic or ephemeral ports, 49152 to 65535, are handed out temporarily to client connections. This reference focuses on the well-known and the commonly used registered ports.

What is the difference between a TCP and a UDP port?

TCP and UDP are two transport protocols, and a port number can be used by either independently. TCP is connection-based and reliable: it sets up a handshake and guarantees ordered, complete delivery, which suits web pages, email, and databases. UDP is connectionless and fire-and-forget, trading reliability for low latency, which suits DNS lookups, video calls, and time sync. A few services, such as DNS on port 53, use both - UDP for quick queries and TCP for larger transfers.

Why is a port "already in use"?

Only one process can listen on a given port and address at a time, so if a second program tries to bind the same port the operating system refuses with an "address already in use" error. It usually means an earlier instance is still running, or another app grabbed the port first. Find the culprit with lsof -i :PORT or ss -tulpn, then stop it or start your service on a different port. This is why dev servers so often fall back to 3000, 5173, or 8080.

Can I run a service on any port I like?

Largely yes, with two caveats. Binding a port under 1024 normally requires administrator or root privileges, which is why development servers default to higher numbers. And the port has to be free - not already taken by another process. Beyond that you are free to choose, though sticking to the conventional number for a service makes life easier for clients and firewalls that expect it. Nothing about the number itself is magic; it is just an agreed label.